!!!!! ! ! !!!!!!!!!!

!

!!!Accelerator Division !!!!!!!!!!

Lali Tchelidze

General Overview of Beam Loss Monitoring Systems

1 April 2011

ESS AD Technical Note ESS/AD/0014

!

A technical report on

GENERAL OVERVIEW OF BEAM LOSS

MONITORING SYSTEMS

by

Lali Tchelidze∗

European Spallation Source ESS AB

P.O. Box 176, SE-221 00 Lund, Sweden

∗Contact: lali.tchelidze@esss.se

Abstract

Beam loss monitors are typically designed to measure the amount and

location of beam losses. Also, they are used to tune the accelerator and

detect/localize problems in beam steering. Beam loss monitors are a very

important part of the machine protection system and therefore need to be

accurately designed and reliably constructed. This report briefly describes

proton interaction mechanisms with matter, the radiation effects of the inter-

action and the ways to detect the radiation. The types of radiation detectors

that can be used for measuring beam losses are described and analysis for se-

lection is performed (with an emphasis on using the loss monitors for proton

linear accelerators). Also the plan on how to optimize the system to pre-

cisely detect and distinguish between fast (occurring within microseconds)

and slow (occurring within seconds) losses is given.

Contents

1 Introduction to Beam Loss Monitors 2

2 Interaction of Protons with Matter 4

2.1 Overview of the interaction processes . . . . . . . . . . . . . . 4

2.1.1 Electromagnetic interactions . . . . . . . . . . . . . . 7

2.1.2 Strong interactions . . . . . . . . . . . . . . . . . . . . 8

2.2 Secondary particle production . . . . . . . . . . . . . . . . . . 9

2.3 Interactions of secondary particles . . . . . . . . . . . . . . . 11

3 Beam Loss Detection 14

3.1 Scintillator detectors . . . . . . . . . . . . . . . . . . . . . . . 14

3.2 Ionization detectors . . . . . . . . . . . . . . . . . . . . . . . 15

3.2.1 Response time of ionization chamber . . . . . . . . . . 17

3.2.2 Dynamic range of ionization chamber . . . . . . . . . 18

3.2.3 Calibration and sensitivity of ionization chamber . . . 18

3.2.4 PIN diode . . . . . . . . . . . . . . . . . . . . . . . . . 19

3.2.5 Diamond detectors . . . . . . . . . . . . . . . . . . . . 19

3.3 Comparison of different beam loss monitors . . . . . . . . . . 20

4 Beam loss monitors as a part of Machine Protection System 22

1

Chapter 1

Introduction to Beam Loss

Monitors

A number of particle accelerators exist to accelerate charged particles to

high speeds for different purposes. The transmission of the charged particles

from the source to the target is never 100%. Fraction of the beam is lost

during the steering and the losses must be carefully controlled to achive

effective transmission. Also, the lost particles cause activation of accelerator

components. The radiation as well as heating due to particle’s lost energy

might damage the accelerator or surrounded instruments. So called Beam

Loss Monitors (BLM) are used to detect secondary reaction products and

thus control losses. At every high current accelerator facitily these monitors

are installed for the protection of accelerator components, environment and

personnel. Beam loss monitors are sensitive devices that can detect very low

losses, therefore they are also used in machine tunning.

A number of different types of BLMs exist. A careful analysis of the

location and time structure of losses has to be performed before choosing

suitable detection devices. Beam loss monitors should be able to measure

irregular/uncontrolled/fast losses and regular/controlled/slow losses. The

irregular losses are usually avoidable and are a result of a misaligned beam

2

or of a faulty condition of accelerator elements, vacuum problems, etc. This

type of losses can damage the beam pipe, activate the environment, dam-

age the equipment in the accelerator tunnel, or quench superconducting

elements. Thus, a fast response is required. Regular beam losses are not

avoidable and are a result of aperture limits on the collimator system, beam

size, residual gas scattering, not ideal beam alignment, instabilities in the

beam, halo scraping, etc. The regular beam losses need to be monitored

constantly and kept low for allowing hands-on maintenance on the machine.

The most common choice for beam loss detection and monitoring, for

hadron accelerators, is ionization chamber. We will describe it’s working

principle, together with other types of detectors in later chapters, but here

we will list the main properties of ionization chamber, which make them

the best choice for hadron machines (we will as well point out some of their

limitations). Ionization chambers are easy to assemble and operate. They

are relatively easy to calibrate and are very radiation hard. They also are

quite sensitive and a have high dynamic range. All these properties make

them robust and easy solution for controlling the losses. The limitations are

response time (not fast enough to resolve the micro structure of the beam

pulse) and sensitivity for low energy part of the accelerator. Fast scintillators

or neutron detectors can replace ionization chambers when faster response

is needed or when the low energ losses need to be measured.

3

Chapter 2

Interaction of Protons with

Matter

In this chapter proton–matter interaction processes leading to secondary par-

ticle generation is reviewed briefly.

2.1 Overview of the interaction processes

Protons are positively charged particles and interact with matter by elec-

tromagnetic and strong interactions.

• Electromagnetic interaction is the main interaction process taking

place for lower energies (≤ 100 MeV). This kind of interaction can

occur with the electric field of the nuclei or with the atomic electons

of the nuclei.

Interaction with the electric field is elastic scattering and results in

small change in the direction of protons motion. The effect of each

small scattering can accumulate and a phenomenon called multiple

coulomb scattering can occur.

4

The interaction with atomic electrons are generally inelastic. This

means that protons lose some amount of energy by exciting orbital

electrons, or by ionizing atoms (kicking electrons out of their orbit).

These generally do not change the direction of the motion of a pro-

ton, but many scatters do reduse the energy of it. Both, ionization

process and excitation are statistical processes and two identical par-

ticles will not create the same number of electron-ion pairs. Because

of this, the number of pairs produced does not equal to the energy loss

divided by the ionization potential (some energy is lost on excitat-

ing electrons), but to the energy loss divided by the mean energy for

ion-electron pair creation (W parameter). So, one should distinguish

between excitation potential, ionization potential and the W parame-

ter. The excitation potential is the energy needed to excite an atom

from it’s ground state and there is first, second, third, etc. excitation

potentials. The ionization potential is the energy needed to remove

electron from an atom to an infitine distance. And the W parameter

is the mean energy needed for one electron-ion pair creation. The W

parameter for most gases is around 30 eV per electron-ion peir. Once

again, these processes, excitation and ionization, account for the most

of the change of the proton’s energy.

• Stong interaction range is short (of order of fermi) and protons interact

with other protons and neutrons in a nucleus by colliding with them.

The probability of these colissions is given by the cross section of the

interaction. Proton-nucleus interaction via strong interaction can be

either elastic or inelastic scattering.

If the interaction is elastic, the proton scatters at some angle, retaining

momentum and identitiy.

If the interaction is inelastic, the proton get “absorbed” in the in-

5

teraction. That is most of it’s energy is transferred to break up the

nucleus, in the process producing subatomic particles called pions.

If the energy of protons is sufficient (≥ 100 MeV), spallation hap-

pens. Typically spallation process produces large number of neutrons

with energy from a few to a few tens of MeV. 1 GeV proton interact-

ing with a heavy matter produces around 30 thermal neutrons, thus

about 30 MeV is required to produce one external thermal neutron. It

was observed, experimentally, that the neutron yield is proportional

to the beam power, independent of the beam energy. For example, a

2 GeV, 1 mA average beam current proton beam produces the same

nutron flux as a 1 GeV, 2 mA beam. This is the reason for discussing

multi-MW accelerators rather than high-energy machines.

Summary of the interaction processes and some of the secondary particles

generated is shown in Figure 2.1. Each process has a certain probability,

which is higly dependent on the energy of the incident proton.

Figure 2.1: The overview of the proton interaction processes with matter

and a list of some of the secondary particles produced. The processes are of

statistical nature and occur with a certain probability each.

6

Head-on collisions of protons onto electrons can happen as well, but the

percentage of energy loss on this interactions is very low (< 3% for 1 GeV

protons), except for some extreme relativistic cases. The maximum energy

loss per collision is

Qmax =4meMpEp

(me +Mp)2. (2.1)

2.1.1 Electromagnetic interactions

The dominant energy loss mechanism, through electromagnetic interactions,

is ionization. The average kinetic energy loss per unit distance in the tar-

get material (or sometimes refered to as an average stopping power of the

medium for the particle of interest), by electromagnetic interactions, with a

good approximation is described by the Bethe-Bloch equation:

−dE

dx=

4π

mec2· nz

2

β2·�

e2

4π�0

�2 �ln

�2β2mc2

I

�− ln

�1− β2

�− β2

�, (2.2)

where me = electron mass, Zp = ze, v = βc and E are projectile charge,

velocity and energy respectively, e = electron charge, c = the speed of light,

x = distance traveled by the projectile particle, I = the mean ionization

potential and n = ρNAZAMu

is electron density of target; NA = Avogadro’s

number, A and Z are mass and atomic numbers of the target and ρ and Mu

are density and molar mass of the target.

This is an average linear rate of energy loss and is a quantity of a great

importance. With a good approximation I(Z) = 16 eV ·Z0.9 and the stop-

ping range by electro-magnetic processes can be given as:

Rel(E) =E1.75

1.75 cR, (2.3)

where cR ≈ 500ρZA · z · A0.75

p , and Ap is an atomic number of the projec-

tile. Protons travel a few centimeters to tens of centimeters through matter

7

before they stop by interactions with the object either through strong or

electromagnetic forces. These distances are sometimes also called interac-

tion or attenuation length.

Below is shown a total stopping power of a proton in iron together with

electromagnetic and nuclear stopping powers and ranges as a function of

proton’s energy (in the energy range or 0 - 1 GeV).

Figure 2.2: Proton stopping power in iron as a function of it’s energy

It is convenient to introduce MIP = Minimum Ionizing Particle which

is a particle whose mean energy loss rate through matter is close to the

minimume. MIP ≈ 1 - 2 MeV-cm2/gram for most materials. We will revisit

this concept when describing the ionization chambers in the next chapters.

2.1.2 Strong interactions

Protons may undergo nuclear reactions when travelling through a matter.

For protons with energy 100 MeV or more, spallation is the most probably

reaction to occur. The process is a source of neutrons. The reaction cross

section can be described with:

8

Figure 2.3: Proton range in iron as a function of it’s energy

σ(E) = σ0

�1− Vc

E

�, (2.4)

with

σ0 =π

100

�1.3 A1/3 + 1

�2barns, (2.5)

and Vc coulomb barrier

Vc =1.44 Z

1.3 A1/3 + 1

A+Ap

A. (2.6)

From above it is possible to derive the nuclear range.

Rnuc(E) =A

0.6ρσ0=

31 A1/3

ρ. (2.7)

2.2 Secondary particle production

When a high energy proton hits the vacuum beam pipe and other surrounded

objects, the processes described in previous sections occur and a number of

9

secondary particles are generated. Some of these secondary particles are

electrons, pions, neutrons, and photons. Other elementary particles can be

produced via various other (generally less probable) nuclear reactions. For

proton energies above 1 GeV, the probability that a nuclear interaction will

happen rises to nearly 100% and the neutron yield scales approximately with

kinetic energies above 1 GeV (see Figure 2.4). In addition, hadron showers

are possible resulting in various species of elementary particles.

Figure 2.4: Left: probability of proton inelastic nuclear reaction within its

range, and proton range as a function of energy. Right: Neutron yield for

various metals as a function of proton’s energy.

Charged particles are relatively quicly stopped by the surrounded ma-

terial, but the nuetrons produced can travel long distances. Except the

radioactive nuclei production all other processes are fast (faster than 10 ns).

10

Due to the kinematics of the primary interaction, the secondary particle

emission shower is mostly forward peaked (not for low energies). Some of

the interaction processes that the secondary particles experience are listed

and described briefly in the following section.

2.3 Interactions of secondary particles

Electrons (or positrons):

Electrons, as other charged particles undergo multiple scattering and

also electromagnetic stopping when passing through a matter (already de-

scribed above). However, since they are light particles, there is another

stopping mechanism: emission of electromagnetic radiation produced from

scattering in the electric field of the nucleus. This is called a bremsstrahlung

radiation and is classicaly understood as a braking radiation. At energies of

a few MeV, this is still relatively small factor. For these energies, ionization

loss dominates, resulting in soft X-rays, which are attenuated within short

distances. However, for larger energies energy loss due to bremsstrahlung

radiation gets comparable to the losses by collisions/ionization. The total

energy loss is a sum of a radiation and a collision/ionization energy losses.

�dE

dx

�

tot

=

�dE

dx

�

brems

+

�dE

dx

�

coll/ion

. (2.8)

Photons:

Photons interact completely different way with the matter compared to

charged particles. The main interaction processes are: photo-electric effect,

compton scattering and pair-production. Also, it is possible, but not as

common, to have nuclear reactions (like (γ, n), etc. reactions). In photo-

electric effect, the photon gets absorbed by an atom, releasing electron. The

two differences compared to charged particle interactions is that photons

penetrate much deeper in a matter and the photon beam is not degraded

11

in energy, only attenuated in intensity. This is because the photons get

removed from the beam completely (if they get absorbed or scattered) and

if they manage to pass through, they usually don’t undergo any interactions

at all and retain their energy.

Photoelectric effect is a process, in which a photon gets absorbed by an

atomic electron and the electron gets ejected with an energy, by binding

energy less than the photon energy.

Compton scattering is a scattering of a photon off of a free electron.

In matter, of course, electrons are bound, but if the photon energy is high

enough, then the binding energy of the electrons can be disregarded and

they can be considered free. During the scattering, photon changes energy

and direction and is refered to as “different” photon which is not a part of

the primary photon beam any more.

The pair-production occurs when the energy of the photon is greated

than the rest energy of electron-positron pair (1.022 MeV), then, the electron-

positron pair is generated.

Obviously, all the above three processes happen in combined manner.

For low energies, the photo-electric effects dominates, in the medium energy

range the compton scattering has the highest probability and for higher

energies, the pair-production takes over.

Neutrons:

Due to lack of electric charge, neutrons don’t experience any Coulomb

interactions. Instead, the principal interaction is a strong interaction with

the nuclei. Depending on energy, neutrons can experience elastic or inelastic

scattering from the nuclei. When inelastically scattered, the nucleus comes

to an excitation and then decays emitting gammas, or other form of ra-

diation. The neutron has to have enough energy to cause the excitation.

Usually the excitations threshold is of order of 1 MeV or more. Below this

threshold, elastic scatterings occur only. Neutron capture is another process

12

that can happen. In general this is led by photon emission and the cross

section of the process is inversely proportional to the velocity of the neutron.

So, this is most common to happen for slow neutrons. Other nuclear reac-

tions, such as (n, p), (n, d), (n,α), (n,αp), etc. can happen also, in which

neutron gets absorbed and charged particle gets emitted. Like the previ-

ous capture (radiative neutron capture), the cross section falls with energy

as 1/v. Fission, (n, f) is possible for thermal energies. For high neutron

energies (> 100 MeV) hadron showers are produced.

13

Chapter 3

Beam Loss Detection

When protons interact with matter a shower of diverse secondary particleas

are produced. Electrons, positrons, neutrons, photons, pions are some of the

possible particles created. All these particles (neutrons are the majority of

the secondary particles for high energy incident primary protons), including

the primary particles (protons) need ot be detected with approproate radia-

tion detectors. Below is described the types of detectors might be used.

The nature of the losses and secondary particle genetarion is very dif-

ferent along a typical proton linear accelerator, depending on the energy of

the protons. There is a number of detectors that can be used for beam loss

monitoring. These are: plastic and liquid scintillator detectors, neutron sen-

sitive scintillators, ionization chambers, secondary electron monitors, PIN

diodes, diamond detectors, optical fibers, etc. Brief overview of the working

mechanism of a few of them is given below.

3.1 Scintillator detectors

Scintillator detectors are one of the type of detectors widely used in nuclear

and particle physics. It’s working mechanism is based on the fact that when

14

struck by the radiation small flashes of lights are generated, i.e. scintillations

happens. Scintillator is usually optically coupled with an amplifier devise,

usually a photomultiplier. Coupling is done directly or via a light guide.

When the light is transformed to the photomultiplier (PMT) a weak electric

signal is generated with the photo-electrons. This signal is then amplified

by the electro amplifying mechanism.

Below is shown a schematic diagram of a scintillator detector.

Figure 3.1: Schematic diagram of a scintillator detector.

In principle, scintillator responds to any kind of radiation which can

directly or indirectly excite the molecules of it and produce light. However,

for a given type of radiation, a proper scintillator has to be chosen to get a

useful signal.

Plastic scintillators detect charged particles due to their electronic stop-

ping. They are also sensitive to neutrons due to their scattering on the

hydrogen atoms of the polymers. Scintillators are very useful tool for pho-

ton detection as well.

3.2 Ionization detectors

Ionization detectors were the first radiation detection devices developed.

These instruments are based on collection of the ionization charge, produced

in a gas by passing radiation. Because of greater mobility of electrons and

ions in gases, gaseous ionization chambers are the most commonly used ones.

15

The basic principle is demonstrated in Figure 3.2. Ionization chambers are

utilised to measure the fluence of charged particles and are very staright

forward to use, because the induced current is proportional to the incident

particle flux.

Figure 3.2: Ionization chamber working mechanism.

If there was no voltage applied between the electrodes the produced

electron-ion pairs would recombine back and there would be no signal ob-

served. By applying an increasing voltage, the fraction of the charge col-

lected will be increased, reaching the saturation point, at which all the

produced charges are collected. The region where ionization charges get

completely collected is called an ionization chamber region (see Figure 3.3).

For higher voltages, the electrons and ions themselves ionize the gas and

more charges get collected than produced by the primary particles of inter-

est. This region is called a proportional region. We need to operate in the

ionization chamber region.

For many reasons, ionization chambers are the a primary tool for beam

loss detection in hadron accelerators. A cylindrical ionization chamber is

of a common use. The cylindrical electrodes inclose a gas with a certain

effective volume that determines the detectors sensitivity (see Figure 3.4).

16

Figure 3.3: Different regions of operation of gas-filles detector.

Figure 3.4: Ionization chamber working mechanism.

3.2.1 Response time of ionization chamber

The transit time of the ions through a gap of an ionization chamber is given

by:

t =D

vion=

D2

µ0V (P0/P ), (3.1)

where vion = µ0E(P0/P ), µ0 is ion mobility, vion is ion velocity, V is the

applied voltage, D is the gap between the electrodes and P/P0 is the pressure

in units of the atmospheric pressure. For cylindrical geometry:

17

D2 =a2 − b2

2ln�ab

�, (3.2)

a and b are inner and outer radii of the cylindrical electrodes.

3.2.2 Dynamic range of ionization chamber

Dynamic range is defined as the ratio of the largest and smallest value of

radiation that can be measured with a radiation detector. The upper limit

is given by the non-linearity due to the recombination rate at high dose (the

typical current in such a case is a few hundred µA). The lower limit is given

by the dark current between the two electrodes. Dark currents should be

kept at around 10 pA or lower. Careful design to achieve this is needed.

This gives a typical dynamic range of 106 - 1-8. Such high dynamic range

needs special signal treatment (logarithmic amlifiers, high ADC resolutions,

current to frequency converters, etc.).

3.2.3 Calibration and sensitivity of ionization chamber

As already mentioned before, MIP is a minimum ionizing partcle. It is a

particle whose mean energy loss rate through matter is close to the mini-

mume. MIP ≈ 1 - 2 MeV-cm2/gram for most materials. It is convenient

to introduce MIP, because then we can describe the radiation response of

a beam loss monitor in terms of either energy deposition (100 ergs/gram),

or in terms of a charged particle flux (3.1 · 107 MIPs/cm2). This is valid,

because:

1 rad =100 ergs

gram=

100 ergs

gram· MeV

1.6 · 10−6ergs· MIP · gram2 MeV · cm2

=

= 3.1 · 107 MIPs/cm2.

The sensitivity of a beam loss monitor SBLM [C/rad] is calculated for

1 rad dose created by MIPs. The number of electrons ne produced in the

18

gap D of an ionization chamber by one MIP is

ne =D · ρW

· dEdx

, (3.3)

where dEdx is given by Bethe-Bloch formula, ρ is a density of the medium and

W is the W value (energy needed to create on electron-ion pair).

For example, if we have and Argon-filled chamber, then

ρ = 1.661 · 10−3g/cm2 (at 200C, 1 atm),

dE

dx= 1.52 MeV/(g/cm2),

and thus

ne ≈ 100 cm−1 ·D[e/MIP ].

SBLM , normilized to 1 rad depends on it’s size. If we assume a 1 liter

Ar-filled chamber with 100% charge collection efficiency, then

SBLM ≈ 638nC

rad.

3.2.4 PIN diode

PIN diode is essentially a solid state ionization chamber. In the typically

100 µm thick depletion layer of the doped Si-crystal, electron-hole pairs are

generated. For one MIP, about 104 electron-ion pairs are generated. See

Figure 3.5.

3.2.5 Diamond detectors

Diamond detectors, which are solid state ionization chambers are unique for

their fast and efficient charge collection and high radiation tolerance. This

is a proven technology and are used at many labs around Europe, USA and

Japan. The diamond detectors are usually located inside the beam pipe and

are mainly used to measure the beam intensity and beam energy spectrum.

19

Figure 3.5: Sheme of two PIN diodes driven in coincidence.

The tests for using diamond detectors to measure beam losses using beam

halo particles are still ongoing at SPS and LHC.

3.3 Comparison of different beam loss monitors

When choosing the detector, one needs to consider the following properties of

it: intrinsic sensitivity, speed, dynamic range, radiation hardness, how easy

is it to calibration, calibration uniformity, stability, reliability, size and cost.

Obviously, we want something that ideally has a high intrinsic sensitivity, is

fast, has a very high dynamic range, is very radiation hard, easy to calibrate,

the calibration is uniform from unit to unit, is stable, reliabl, small in size

and cheap.

For hadron accelerators, the ion chambers are the best choice. However

they have limitations. They are not sensitive enough for low energies, and

they are not fast enough to resolve micro-pulse structure (when that is of

importance). In both cases scintillator detectors can be used. However,

they require excessive recalibration and they degrade in radiation environ-

ment rapidly. But their sensitivity can be adjusted highly and they can

detect the losses in low energy region of accelerator. Scintillators are faster

20

too (typically they have of about 1 ns responce time compared to typical

ionization chamber responce time of about 1 µs) and can record the micro-

structure of the pulse. Nuetron sensitive scintillator detectors can also be

used for loss detection, but the losses can’t be easily localized at all.

21

Chapter 4

Beam loss monitors as a part

of Machine Protection

System

A low energy and low intensity beams in accelerators can’t cause any dam-

age even if it’s all lost in the vacuum chamber. However, in high energy and

high intensity accelerators beam losses can be of a great issue. Losses local-

ized in a very small volumes can burn holes in the accelerator components,

quench the superconducting magnets, or damage other equipment around

the accelerator.

Although, beam loss monitors are instruments that are primarily de-

signed to measure the beam losses, they are also frequently used for ma-

chine protection. By proper calibration, BLM signals are used to trigger the

beam dump as soon as the energy deposition in the magnets or in the vac-

uum pipe reaches the levels set to avoid quenches and other possible harms.

The thresholds are typicelly set based on experience. They have to be not

too low not to prevent continuous operation of a machine.

To set the trigger levels allowed losses must be known. This depends on

22

the perticle energy and the duration of the loss. Thus, the loss signal needs

to be integrated over few time periods and the threshold needs to be set for

each of them.

BLMs are used to monitor the activation levels due to losses. Losses lead

to nuclear reactions, which procude radioactive nucli with lifetime of hours

or more. This prevents access to the accelerator tunnel. As a general rule,

to allow the maintenance on the accelerator, the power dissipation of the

beam losses should be lower than 1 W/m. This corresponds to 0.1 rad/hour

at 30 cm from the surface of the accelerator, after 4 hous of cool-down.

The position of beam loss monitors should be well chosen. Simulation

tools (FLUKA, Geant4, MARS) are used to calculate the locations of the

highest particle fluxes. Of course, an assumption of the primary particle

losses is needed. Since the beam is largest at the quadrupole magnets, these

are usually the locations where the BLMs need to placed.

23